1 + 1<2? Unveiling the impact of AI-assisted disclosure on service satisfaction in professional services

Abstract

AI-assisted human employees have become widespread, particularly in professional services such as healthcare, law, and management consulting. However, little is known about how consumers respond to AI-assisted employees in professional services. This study examines how disclosing AI-assisted employees in professional services impacts service satisfaction and explores the underlying mechanisms and boundary conditions. Through six experiments (N = 1664), we find that disclosing (vs. non-disclosure) AI-assisted employees leads consumers to infer that the employees possess lower professional competence, negatively affecting satisfaction. This effect disappears when consumers perceive AI to have high capability. Moreover, the negative impact is mitigated when a complementary capability or self-selection disclosure strategy is employed. This research provides new insights into the impact of human-AI collaboration and AI disclosure in service delivery. The findings also offer practical implications for organizations and employees using AI assistance in professional service.